T639模式20 h降水预报产品对和田地区降水量预报准确率的检验
Verification of the Accuracy of the 20 h Precipitation Forecast Product of T639 Model to the Precipitation Forecast in Hotan Area
摘要: 利用2014年1~12月和田地区皮山(51818)、和田(51828)、民丰(51839)三个站点降水量级进行检验,得出以下结论:1) 一般降水预报上,皮山(51818)和和田(51828) T639模式预报1~12月准确率较低,在10%以下,民丰(51839) 1~12月准确率较低,在3%以下;2) 各量级降水预报准确率上,皮山(51818) T639模式预报12~72 h对小量降水预报准确率为5%~11%,空报率较大,达91%,和田(51828) 12~72 h对小量降水预报准确率为10%,空报率较大,达84%,民丰(51839) 12~72 h对小量降水预报准确率为5%,空报率较大,达94%;3) 通过检验三个站点一年的降水资料发现,虽然T639模式预报产品已有很大发展,但降水预报准确率较低,在3%~10%,空报率较大,在81%~98%;4) 2014年,和田地区为降水偏少的干旱年份,年内只出现了2次较明显的天气过程,出现在夏季的6月18~24日和7月1~19日,这两次降水期间T639模式预报的降水预报准确率在15%~21%,其余时段全地区基本无降水,但模式预报空报率仍大,导致全年预报准确的下滑。
Abstract: The precipitation magnitude at Pishan (51818), Hotan (51828) and Minfeng (51839) stations in Hotan region from January to December 2014 was used for inspection. The following conclusions are drawn: 1) In general precipitation forecast, the accuracy of Pishan (51818) and Hotan (51828) T639 models is low from January to December, below 10%, Minfeng (51839) has a low accuracy rate from January to December, below 3%; 2) On the accuracy rate of precipitation prediction of all orders of magnitude, the accuracy of Pishan (51818) T639 model forecast for small amount of precipitation in 12~72 h is 5%~11%, and the false forecast rate is 91%. Hotan (51828) 12~72 h accuracy rate of small precipitation forecast is 10%, and the rate of empty forecast is relatively high, reaching 84%. Minfeng (51839) has the accuracy rate of small precipitation forecast for 12~72 h 5%, with a large vacancy rate of 94%; 3) Through the examination of the one-year precipitation data of three stations, it is found that although the T639 model prediction products have great development, but the accuracy rate of precipitation forecast is relatively low, ranging from 3% to 10%, and the rate of air forecast is relatively high, ranging from 81% to 98%; 4) In 2014, Hotan is a drought year with less precipitation. There are only two obvious weather processes in the year, which occur on June 18~24 and July 1~19 in summer, during the two precipitation periods, the accuracy rate of precipitation forecast by T639 model is 15%~21%, and there is basically no precipitation in the whole region in other periods, but the false reporting rate is still high, leading to the decline of the accuracy of the annual forecast.
文章引用:阿依夏木古丽•买买提, 阿依夏木•买买提托合提, 古扎丽努尔•艾尼瓦尔. T639模式20 h降水预报产品对和田地区降水量预报准确率的检验[J]. 气候变化研究快报, 2022, 11(5): 876-887. https://doi.org/10.12677/CCRL.2022.115091

参考文献

[1] 皇甫雪官. 国家气象中心集合数值预报检验评价[J]. 应用气象学报, 2002, 13(1): 29-36.
[2] 邓海光, 曾小团. 统计预测方法的试验比较[J]. 贵州气象, 2006, 30(1): 26-28.
[3] 牟换, 阿不力米提江•阿布力克木, 赵凤环. T639和德国模式对新疆大降水预报的检验[J]. 沙漠与绿洲气象, 2013, 7(1): 12-15.
[4] 熊传辉, 马安国, 饶传新, 等. 三种数值预报产品在清江流域面雨量预报中的应用检验[J]. 湖北气象, 2004(1): 24-26.
[5] 缪强, 谢瑞国. 数值预报产品释用若干问题的初步分析[J]. 四川气象, 2001(1): 9-12.
[6] 赵守良. 无冷高压的南支低槽引起的雨雪凝冻过程分析[J]. 贵州气象, 2000, 24(2): 11-13.
[7] 赵敏芬, 宋海强, 卢兆民, 等. 淄博市紫外线辐射强度变化特征分析及预报检验[J]. 山东气象, 2005, 25(4): 38-39.
[8] 高建峰, 张红雨, 武捷, 等. 中尺度模式MM5在山西预报业务中的应用[J]. 山西气象, 2002(4): 1-5.
[9] 段靖, 苗春生. 人工神经网络在梅雨期短期降水分级预报中的应用[J]. 气象, 2003, 31(8): 31-36.
[10] 秦贺, 赵克明, 汤浩, 等. 新疆中强天气过程GRAPES区域模式降水预报检验[J]. 沙漠与绿洲气象, 2010, 4(6): 12-15.
[11] 张学文, 张家宝. 新疆气象手册[M]. 北京: 气象出版社, 2006.